On May 22, 1999, the IBM Institute for Advanced Commerce sponsored an informal evening seminar on the topic of gathering statistics about the digital economy. Last year, President Clinton made a commitment that the Federal government would strengthen its monitoring of the benefits of e-commerce. To further that aim, we ran a meeting where senior government officials responsible for policy and data analysis could meet with experts from academia, consulting firms, and other companies.

The problem is acute, since electronic commerce will soon represent a significant percentage of overall business in the United States. a significant fraction of overall business. It is also complex, since different constituencies need different analyses and there is a wide choice of raw data to be gathered and combined.

The meeting addressed these topics:

• What sort of information would be of most use to private sector firms? What segmentation, granularity, frequency, and time lags would be most acceptable?
   
  
  • What sorts of data can be gathered and who would use them? Are data about bits and pages, number of transactions, and dollar volume sufficient? What sorts of information on employment, skill requirements?
     
    
  • What are the appropriate ways to segment the data? Along traditional categories of business? Between business-to-business and business-to-consumer transactions? Between physical goods and information goods? Between financial transfers and security sales and other economic activity?
     
    
  • What retrospective data exist for validation of models and methods?
     
    
  • How do we define precisely the data that are needed? How do we work in a world where new concepts and business categories arise every few months?
     
    
• How should the U. S. government collect the information? The Census is already moving to capture more information more often about e-commerce firms. What else needs to be done to get an accurate picture of the rapidly evolving digital economy?
   
  
  • How do differences between the digital economy and the traditional economy affect the sorts of data that can be collected, analyzed, and modeled?
     
    
  • What are the possible technical and business sources of data? How rapidly does the set of sources change, how often should reports be requested, and what sorts of sampling can be done in the future? Are there points in the e-commerce process (such as credit card charges or electronic funds transfers) where we could reliably and inexpensively get reliable summary data?
     
    
  • What precisely are the definitions of information to be gathered, in order to identify business segments and uses, and to distinguish processes that are traditional and those that are to be include in the digital economy?
     
    
• How can the Federal government evaluate the impact of e-commerce on employment, productivity, domestic product, and trade balances? What analytic tools are available or are needed? What models exist or can be constructed?
   
  
• How can the government work with the private sector to mutual advantage to take advantage of their different strengths and requirements?
   
  

Among the attendees were economists and statisticians from major federal government data-gathering organizations (Bureau of Labor Statistics, Bureau of Economic Affairs, Census Bureau, National Science Foundation) as well as policy makers from the National Economics Council and the Department of Commerce. In addition, there were representatives of the OECD, high-level analysts from a half dozen IT consulting firms, and senior executives from large companies, and professors from three universities.

The meeting began with short comments from Stuart Feldman (IBM Institute for Advanced Commerce), Tom Kalil (National Economics Council), Elliot Maxwell (Department of Commerce) and Brian Kahin (OSTP). The discussion on the types of information that would be most valuable was led by William Archey (President of the American Electronic Association). The discussion on how the federal government could collect data was led by Frederick Knickerbocker (Chief Economist of the U. S. Census Bureau).

This meeting opened the discussion between some of the key experts in the public and private sectors. In order to meet President Clinton's commitment, much work will need to be done to define the information. that will be needed, to choose the best methods to gather it, and to make best use of it for public policy and corporate planning. We expect to organize future meetings on specific technical issues raised during this seminar.

###